Welded spin holes in ellipsoidal head jacketed vessels

[digger02]

I am operating a large number of jacketted vessels at 1425F, with the inner vessel containing a lethal substance. These vessels have an inner lining of mild steel (.75&quot, and an outer vessel being T316L (0.5&quot. There is a gap between the jackets ranging from 1/8" to 1/4". The ends of the vessel are dished ellipsoidal heads that were formed on a spindled roll machine. The spin holes are sometimes 1", sometimes 1.5", and very rarely around 2". We have found what appears to be a process change at the fabricator. On our older vessels, the mild steel spin hole and stainless steel spin hole were filled individually and seperately, although filled entirely with weld filler metal. They appear to have used fluxed core welding rods, as there was some remnant slag/flux visible between the two "filler plugs" as I'll call them. Even with this design, the inner mild steel filler plug would fail and leak a lethal molten salt into the gap between the vessels, but usually after a high number of hours of service. This leak is not catostrophic, as the mild steel holds the molten salt for a long enough period of time that we usually catch the failure before the a leak to the outside of the stainless steel (atmosphere) (we see stuff in our process data).

Question #1 - Any obvious violations in the Pressure Vessel code here ?


Now, for the process change. Our latest batch of vessels from the fabricator have been found to have a one piece plug which fills both the spin hole in the inner mild steel and the outer stainless jacket. Again, the fill was done with all filler metal, and no wrought plug. The welder started laying in flux core wire on the inner mild steel pot. He then went to the outside of the stainless vessel, and using the mild steel weld metal as his backing, began laying in T309 filler metal (again, with flux core wire). As he got nearer to the outside of the stainless vessel, he switched to T316L flux core weld rods, and layed in filler metal to he had a blob sticking out of the outer stainless vessel. He then used a die grinder to smooth out both the stainless side and mild steel sides of the filler-metal plug. We "visibly" inspected the welds after grinding.

Question #2 : Do you see violations of the pressure vessel code here


There's a couple of things that come to mind - visible inspection is not sufficient - it should have been dye penetrant tested. Now that the filler metal plug is connecting the inner vessel to the stainless jacket, it meets the criteria of a welded stay. Does a welded stay have more rules that apply. The intermixing of weld metals has been proven to produce some very low chrome/nickel dilutions, such that brittle phases were formed (martensite). Are there hard rules about welding stainless to common steel to minimize dilution and prevent brittle phase formation ? Another problem is that the fabricator did not account for the stainless jacket expanding more at 1425F than the mild steel, and this resulted in stress buildup on the plug. There was also trapped air in the dished heads which expanded at 1425F. Both of the issues lead to stress applied to the weld-metal filler plug. In some cases, the plug was yanked out of the mild steel spin hole, in other cases the plug yanked out of the stainless spin hole, in some cases the plug fractured horizontally, and some cases are mixes of all three. We had a few leakages of the lethal molten salt, but no one was hurt, and no serious external damage occurred, although some vessels are trashed.

Any thoughts appreciated. Thanks.

Dave

 

[prex]

In my country we say that the world is beautiful because it is varied!
The procedure you outline is IMHO far from good code practice.
For the least a procedure qualification record for the actual sizes, thicknesses, weld metal, etc. should be done.
The best procedure is anyway to use a plug with the same material as head wall (same thickness or higher thickness) and treat the weld of the plug to the wall as a cateogory A joint in a head or as a nozzle weld (category D) and inspect it accordingly to code.
When a single plug for both heads is done as you described, this is really out of question, as you cannot inspect the inside of the weld in the interspace.
By the way it is not clear in your description whether the inner head is really a kind of liner, or is a pressure retaining wall (this depends on how the vessel is calculated).
When the joining of the two heads at the hole is done correctly, thermal stresses or stresses due to air pressure build up shouldn't be a concern as far as the stay is concerned, but is obviously of concern for head stresses! Anyway air should be vented and not allowed to stay in the interspace (you lose also in thermal efficiency). Also pressure build up in the interspace could cause the buckling of the inner head.
What I think should be done is to use a stay in a material that you consider suitable for the process, weld it first to the inner head by a weld from both sides (and a full penetration weld for the best, as you have thermal stresses), then weld it to the outer head from outside, but again, with strong preference, with a full penetration weld (first pass TIG).
Of course you can also invert the above, and weld before to the outer head (double weld) and then to the inner one (I prefer the former, as the inner weld is in contact with the process fluid).
If you can accept a nozzle in that position then make the same using a piece of pipe: you could find this even useful!

Oops! In the above I forgot that you mentioned a lethal substance: you are really playing with a bomb, and I strongly suggest you to hire an expert for this problem!

Anyway for lethal substance a heat treatment is required, so, if the inner wall is a pressure retaining wall, you could be obliged to heat treat it separately from the jacket (but after the plug is welded).
Also full penetration welds are a must, and, if you choose to separately plug the two heads with a plate disc (I suggest to do this), then the weld of the pressure retaining wall (be it the inner or the outer one) needs RX (as the rest of the joints in the vessel). Possibly more issues need be considered: I insist, the full design needs a revision by a professional.
prex

http://www.xcalcs.com

Online tools for structural design